The Royal Society's last conference on risk in 1992 was notable for the squabble between pure and social scientists. Julia Hinde looks at why those involved failed to agree and asks if the prospects for meaningful dialogue are better this time.

Complete decorum reigned until near the end when a psychologist got up from the floor. When he asked that the term 'social construction of risk' be eliminated from the discussion, shouting, clapping and hissing broke out and the meeting was adjourned."

The year was 1992; the unruly event the launch by the Royal Society - the rather conservative bastion of British science - of a report which tackled the sixmillion-dollar question; how do you calculate the risks attached to so many phenomena of modern living, from dodgy beef to nuclear waste dumps?

According to its terms of reference, the report sought to bridge the gap between supposedly objective risk, derived by scientists from statistics, and admittedly subjective risk, as perceived by the public. It failed. Engineers still saw busy roads with low death tolls as objectively safe and the public's insistence on their danger as unsubstantiated. The social scientists, on the other hand, maddened the pure scientists by insisting that objective and subjective risk could not be separated, with some taking the position that risks are culturally constructed - created by people according to their experience and perception of the world.

³ÉÈËVRÊÓƵ

As British anthropologist Mary Douglas describes above, the launch meeting degenerated into a bitter squabble. "I have never been to a launch party before," says John Adams, reader in geography at University College, London, "where the authors of part of the report were being rude about the authors of the other parts."

Rather than being a collaborative discussion of risk, the 1992 report consists of two sections, written from different viewpoints, which in parts flatly contradict one another. Chapters one to four are the work of engineers, statisticians and natural scientists. They cling to the notion that objective and perceived risk are distinct, and that the public's perception of risk is flawed. To improve risk assessment what is needed, say the scientists, is more research and more statistics to improve the accuracy of models.

³ÉÈËVRÊÓƵ

Chapter five, however, where the social scientists take over, starts: "The view that a separation can be maintained between 'objective' and 'subjective' or perceived risk has come under increasing attack, to the extent that it is no longer a mainstream position." In their recent book Accident and Design, social scientists Christopher Hood and David Jones, contributors to the concluding chapters of the 1992 Royal Society report, refer to it as "four chapters good, two chapters bad (with apologies to George Orwell)".

The social scientists suggest that disparate groups view risk in different ways and react to it differently, partly changing the probability of that risk occurring. Dr Adams cites slipping on ice as an example of how different groups may react to risk in their own ways. A game for young children, but a potentially fatal accident for the elderly, Adams says the probability of slipping on ice is influenced by whether someone sees it as fun or dangerous. Old people see the risk of slipping on ice as high, and therefore try to avoid it, thus reducing the probability of a fall, while young children see sliding on ice as fun and jump at the chance. The two groups see the world differently and therefore behave differently, Adams suggests. Women tend to fear all risks more than men and non-whites in America are more fearful than whites. In fact, white men have been shown to be less fearful of most risks than any other grouping.

Britain's roads provide another example that highlights the gulf between the approach to risk of scientists and social scientists. Measured by fatality statistics, they are now more than twice as safe for children than they were 70 years ago. Despite the increase in cars, the road accident death rate for British children is less than half that in the 1920s, the roads are safer now than ever before. But according to many lay people, who remember the time when cars were few and far between, this is certainly not the case.

The Department of Transport belongs to the scientists' camp. It measures the safety of roads by casualty figures alone. It draws a clear line between actual danger and the public's perception of risk and is only prepared to spend money to relieve actual danger. If a road does not have a fatality rate significantly above "normal", funds will not be made available to reduce the alleged problem, however persistently the public complains.

In 1991, Sir Patrick Brown, permanent secretary of Britain's DoT, said that "funds for traffic calming will be judged on casualty savings, not environmental improvements or anxiety relief". People up and down the country might live next to roads they perceive to be dangerous, but without the blood to prove it, the DoT classifies the roads as safe and the anxiety as subjective and emotional.

³ÉÈËVRÊÓƵ

In safety literature generally, says Adams, the mainstream position is still that casualty statistics provide the only reliable measure of safety schemes. Casualty figures may be down, but so is children's exposure to traffic. No longer are parents happy to allow their children to play outside or to walk to school on their own. They keep their children off the streets because they instinctively know they are dangerous. People respond to a proposed threat by being careful, in turn reducing the measurable risk.

One extreme position within the social science camp is that first advocated by Mary Douglas, who argued that perceptions of risk vary with the kind of society or group in which an individual lives. If a risk threatens institutional arrangements that are highly valued by society then fear of that risk will be correspondingly greater.

Next week's attempt by the Royal Society to return to the fraught risk debate might be seen as a dangerous exercise in its own right, undertaken as it is just five years after the 1992 report at a time when public confidence in the objectivity of scientists is at an all time low, rocked by the handling of numerous food scares, from mad cow disease to the recent fatal E.Coli outbreak in Scotland.

³ÉÈËVRÊÓƵ

Organiser of the meeting, John Ashworth, chief executive of the British Library, who was influential in 1992 in involving the social scientists, says renewed discussion is essential. "There is clearly a range of perspectives in this field," he said. "In 1992 the idea was to produce a homogenous report. It did not prove possible. This meeting is a response to that. There needs to be a dialogue between social scientists and pure scientists. Both at least need to know why what the other one does is seen as contentious. There has been a dialogue of the deaf until now."

With just one day for debate and individual contributions restricted to 15 minutes, there is concern that the meeting will merely rehash former arguments and do little to bridge the gaping gulf between the groups. Ragnar Loftstedt, of the University of Surrey, says: "The pure and social scientists don't communicate with one another. They never have and they never will. The engineers cannot stomach the social scientists."

Much faith is being put in Adams, the geographer chosen to start the debate, who has credibility with both sides, as well as the distinction of having made Lewis Wolpert, chairman of the Royal Society's committee for the public understanding of science, reexamine his "deeply held prejudices" and admit in an article the validity of the notion of "risk compensation". This is the theory that people modify their behaviour in response to what they see as altered risks to themselves. Since the introduction of seat-belt laws, for example, the death rate from car accidents has not fallen as expected. This, says Adams, is due to people feeling safer and therefore compensating by driving proportionately less carefully.

Adams believes there is scope for a range of ideas. He distinguishes between, at one end of the spectrum, risks calculable through science, and at the other "virtual risks". The risk of infectious diseases, for example, has been reduced by the application of objective scientific techniques, but at the other extreme there are risks about which scientists have very little objectively to say. "These are 'virtual risks'," says Adams. "I think scientists are capable of accepting that there are some risk uncertainties about which science has nothing yet conclusive to say. I think in this murky area it is obvious that people respond very differently to what appear on the surface to be the same facts. There is scope for cultural relativism here, explaining why people behave in this area in different ways."

³ÉÈËVRÊÓƵ

The challenge facing the Royal Society is to persuade the scientists and the social scientists of the validity of each other's arguments and to see if common ground can be found.